Is RDF-Export graph data structure format useful for migrating from Logseq-MD to Logseq-DB?

Questions & Help Documentation
1

I posted a general question with an AI-generated answer.

There and in the community FAQ, it was suggested to download an existing example of structured data (Logseq DB graph) for reverse engineering.

The best example I’ve found is the GitHub downloadable source of docs.logseq.com.

logseq/docs: Logseq documentation - Green code button - Download ZIP

I also found what I initially thought was a Logseq plugin for exporting the graph to RDF format. RDF is Rich Data Format, an XML data interchange standard. In fact, it’s not a Logseq plugin; it’s a GitHub action, so maybe less related than I thought.

logseq/rdf-export: Export a logseq graph to RDF

Is this a useful approach for import to Logseq-DB?

Or is the Logseq documentation not a good example to model structured data for Logseq DB on?

By default, this action is configured to export the portion of your Logseq graph that:

  • has class pages with properties type:: [[Class]]

  • has property pages with properties type:: [[Property]]

  • has pages with properties type:: [[X]] where X are pages with type:: [[Class]]

For an example of such a graph, see https://github.com/logseq/docs. If you would like to organize your graph differently, you can configure it.

By the way, looking for the latest version of the documentation, which is purely for Logseq-MD?

Tag 0.10.9

2

Seems the page properties are specific to the RDF-export process, not for use by Logseq at all?

Knowledge Management for Tags / Tag Hierarchies - Feedback / Feature Requests - Logseq

Comments by @cldwalker :

In the meantime, I’m happy to share that generating rdf from logseq graphs is doable with nbb-logseq/examples/linked-data at main · logseq/nbb-logseq · GitHub . With the example, a small portion of Logseq’s docs are able to generate RDF.

…

For examples, see the backlinks at https://docs.logseq.com/#/page/feature and https://docs.logseq.com/#/page/ui%20element

I’ll also mention I’m a linked data beginner so there are probably more correct ways to model some of the relations e.g. more use of sameAs instead of url. For now I’m trying to strike a pragmatic balance with rdf output without sacrificing usability in Logseq